displayName
LLM Cost Optimization
workflowKind
operational
triggerType
scheduled
typicalCadence
weekly
complexity
cross-team
description
Reviews and optimises spend across LLM API providers — analysing per-model
token consumption, identifying prompt-length bloat, evaluating cache-hit
rates, testing cheaper model substitutions for low-criticality tasks,
auditing retry and fallback policies that inflate costs, and projecting
budget burn-rate against forecasts. Produces a cost breakdown dashboard
and actionable savings plan. Excludes model fine-tuning work.